Tag: artificial-intelligence

Agentic Orchestration Pattern: Bulkhead Pattern


    """
Pattern 09.03: Bulkhead
Isolate failures through resource partitioning
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class BulkheadState(TypedDict):
    """State for Bulkhead pattern."""
    input: str
    output: str


def agent_function(state: BulkheadState) -> BulkheadState:
    """Bulkhead agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Bulkhead pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_bulkhead():
    """Create Bulkhead workflow."""
    workflow = StateGraph(BulkheadState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_bulkhead()
    result = workflow.invoke({"input": "Test Bulkhead", "output": ""})
    print(f"\n=== Bulkhead Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Bulkhead Pattern

03 Bulkhead Pattern – Detailed Documentation

Overview

File: 09_error_handling/03_bulkhead.py
Pattern Category: 09 Error Handling
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 09.03: Bulkhead Isolate failures through resource partitioning

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

Purpose of imports: – TypedDict: For type-safe state definitions – StateGraph: LangGraph’s graph-based workflow orchestration – ChatOpenAI: OpenAI LLM integration – HumanMessage: LangChain message formatting

State Definition

class BulkheadState(TypedDict):
    """State for Bulkhead pattern."""
    input: str
    output: str


def agent_function(state: BulkheadState) -> BulkheadState:

BulkheadState: State for Bulkhead pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Bulkhead pattern: Stores prompt = f”Process this using Bulkhead pattern – def create_bulkhead(): Stores def create bulkhead()

Functions

`agent_function()`

Line: 17
Arguments: state

Bulkhead agent implementation.

def agent_function(state: BulkheadState) -> BulkheadState:
    """Bulkhead agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Bulkhead pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_bulkhead():

`create_bulkhead()`

Line: 26
Arguments: “

Create Bulkhead workflow.

def create_bulkhead():
    """Create Bulkhead workflow."""
    workflow = StateGraph(BulkheadState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class BulkheadState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Bulkhead pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: BulkheadState) -> BulkheadState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Bulkhead agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_bulkhead definition

def create_bulkhead():

Defines the create_bulkhead function that processes the workflow state.

Line 27: Module docstring

    """Create Bulkhead workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_bulkhead()
    result = workflow.invoke({"input": "Test Bulkhead", "output": ""})
    print(f"\n=== Bulkhead Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Circuit Breaker Pattern


    """
Pattern 09.02: Circuit Breaker
Prevent cascade failures
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class CircuitBreakerState(TypedDict):
    """State for Circuit Breaker pattern."""
    input: str
    output: str


def agent_function(state: CircuitBreakerState) -> CircuitBreakerState:
    """Circuit Breaker agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Circuit Breaker pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_circuit_breaker():
    """Create Circuit Breaker workflow."""
    workflow = StateGraph(CircuitBreakerState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_circuit_breaker()
    result = workflow.invoke({"input": "Test Circuit Breaker", "output": ""})
    print(f"\n=== Circuit Breaker Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Circuit Breaker Pattern

02 Circuit Breaker Pattern – Detailed Documentation

Overview

File: 09_error_handling/02_circuit_breaker.py
Pattern Category: 09 Error Handling
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 09.02: Circuit Breaker Prevent cascade failures

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class CircuitBreakerState(TypedDict):
    """State for Circuit Breaker pattern."""
    input: str
    output: str


def agent_function(state: CircuitBreakerState) -> CircuitBreakerState:

CircuitBreakerState: State for Circuit Breaker pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Circuit Breaker pattern: Stores prompt = f”Process this using Circuit Breaker pattern – def create_circuit_breaker(): Stores def create circuit breaker()

Functions

`agent_function()`

Line: 17
Arguments: state

Circuit Breaker agent implementation.

def agent_function(state: CircuitBreakerState) -> CircuitBreakerState:
    """Circuit Breaker agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Circuit Breaker pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_circuit_breaker():

`create_circuit_breaker()`

Line: 26
Arguments: “

Create Circuit Breaker workflow.

def create_circuit_breaker():
    """Create Circuit Breaker workflow."""
    workflow = StateGraph(CircuitBreakerState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class CircuitBreakerState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Circuit Breaker pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: CircuitBreakerState) -> CircuitBreakerState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Circuit Breaker agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_circuit_breaker definition

def create_circuit_breaker():

Defines the create_circuit_breaker function that processes the workflow state.

Line 27: Module docstring

    """Create Circuit Breaker workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_circuit_breaker()
    result = workflow.invoke({"input": "Test Circuit Breaker", "output": ""})
    print(f"\n=== Circuit Breaker Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Try-Catch Pattern


    """
Pattern 09.01: Try-Catch
Error detection and handling with recovery
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class TryCatchState(TypedDict):
    """State for Try-Catch pattern."""
    input: str
    output: str


def agent_function(state: TryCatchState) -> TryCatchState:
    """Try-Catch agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Try-Catch pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_try_catch():
    """Create Try-Catch workflow."""
    workflow = StateGraph(TryCatchState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_try_catch()
    result = workflow.invoke({"input": "Test Try-Catch", "output": ""})
    print(f"\n=== Try-Catch Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Try-Catch Pattern

01 Try Catch Pattern – Detailed Documentation

Overview

File: 09_error_handling/01_try_catch.py
Pattern Category: 09 Error Handling
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 09.01: Try-Catch Error detection and handling with recovery

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class TryCatchState(TypedDict):
    """State for Try-Catch pattern."""
    input: str
    output: str


def agent_function(state: TryCatchState) -> TryCatchState:

TryCatchState: State for Try-Catch pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Try-Catch pattern: Stores prompt = f”Process this using Try-Catch pattern – def create_try_catch(): Stores def create try catch()

Functions

`agent_function()`

Line: 17
Arguments: state

Try-Catch agent implementation.

def agent_function(state: TryCatchState) -> TryCatchState:
    """Try-Catch agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Try-Catch pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_try_catch():

`create_try_catch()`

Line: 26
Arguments: “

Create Try-Catch workflow.

def create_try_catch():
    """Create Try-Catch workflow."""
    workflow = StateGraph(TryCatchState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class TryCatchState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Try-Catch pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: TryCatchState) -> TryCatchState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Try-Catch agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_try_catch definition

def create_try_catch():

Defines the create_try_catch function that processes the workflow state.

Line 27: Module docstring

    """Create Try-Catch workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_try_catch()
    result = workflow.invoke({"input": "Test Try-Catch", "output": ""})
    print(f"\n=== Try-Catch Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Blackboard Pattern


    """
Pattern 08.04: Blackboard
Shared knowledge space for collaborative problem solving
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class BlackboardState(TypedDict):
    """State for Blackboard pattern."""
    input: str
    output: str


def agent_function(state: BlackboardState) -> BlackboardState:
    """Blackboard agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Blackboard pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_blackboard():
    """Create Blackboard workflow."""
    workflow = StateGraph(BlackboardState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_blackboard()
    result = workflow.invoke({"input": "Test Blackboard", "output": ""})
    print(f"\n=== Blackboard Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Blackboard Pattern

04 Blackboard Pattern – Detailed Documentation

Overview

File: 08_state_management/04_blackboard.py
Pattern Category: 08 State Management
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 08.04: Blackboard Shared knowledge space for collaborative problem solving

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class BlackboardState(TypedDict):
    """State for Blackboard pattern."""
    input: str
    output: str


def agent_function(state: BlackboardState) -> BlackboardState:

BlackboardState: State for Blackboard pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Blackboard pattern: Stores prompt = f”Process this using Blackboard pattern – def create_blackboard(): Stores def create blackboard()

Functions

`agent_function()`

Line: 17
Arguments: state

Blackboard agent implementation.

def agent_function(state: BlackboardState) -> BlackboardState:
    """Blackboard agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Blackboard pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_blackboard():

`create_blackboard()`

Line: 26
Arguments: “

Create Blackboard workflow.

def create_blackboard():
    """Create Blackboard workflow."""
    workflow = StateGraph(BlackboardState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class BlackboardState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Blackboard pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: BlackboardState) -> BlackboardState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Blackboard agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_blackboard definition

def create_blackboard():

Defines the create_blackboard function that processes the workflow state.

Line 27: Module docstring

    """Create Blackboard workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_blackboard()
    result = workflow.invoke({"input": "Test Blackboard", "output": ""})
    print(f"\n=== Blackboard Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Memory Pattern


    """
Pattern 08.03: Memory
Persistent memory across interactions
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class MemoryState(TypedDict):
    """State for Memory pattern."""
    input: str
    output: str


def agent_function(state: MemoryState) -> MemoryState:
    """Memory agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Memory pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_memory():
    """Create Memory workflow."""
    workflow = StateGraph(MemoryState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_memory()
    result = workflow.invoke({"input": "Test Memory", "output": ""})
    print(f"\n=== Memory Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Memory Pattern

03 Memory Pattern – Detailed Documentation

Overview

File: 08_state_management/03_memory.py
Pattern Category: 08 State Management
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 08.03: Memory Persistent memory across interactions

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class MemoryState(TypedDict):
    """State for Memory pattern."""
    input: str
    output: str


def agent_function(state: MemoryState) -> MemoryState:

MemoryState: State for Memory pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Memory pattern: Stores prompt = f”Process this using Memory pattern – def create_memory(): Stores def create memory()

Functions

`agent_function()`

Line: 17
Arguments: state

Memory agent implementation.

def agent_function(state: MemoryState) -> MemoryState:
    """Memory agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Memory pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_memory():

`create_memory()`

Line: 26
Arguments: “

Create Memory workflow.

def create_memory():
    """Create Memory workflow."""
    workflow = StateGraph(MemoryState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class MemoryState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Memory pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: MemoryState) -> MemoryState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Memory agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_memory definition

def create_memory():

Defines the create_memory function that processes the workflow state.

Line 27: Module docstring

    """Create Memory workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_memory()
    result = workflow.invoke({"input": "Test Memory", "output": ""})
    print(f"\n=== Memory Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Context Passing Pattern


    """
Pattern 08.02: Context Passing
Shared context between agents
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class ContextPassingState(TypedDict):
    """State for Context Passing pattern."""
    input: str
    output: str


def agent_function(state: ContextPassingState) -> ContextPassingState:
    """Context Passing agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Context Passing pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_context_passing():
    """Create Context Passing workflow."""
    workflow = StateGraph(ContextPassingState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_context_passing()
    result = workflow.invoke({"input": "Test Context Passing", "output": ""})
    print(f"\n=== Context Passing Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Context Passing Pattern

02 Context Passing Pattern – Detailed Documentation

Overview

File: 08_state_management/02_context_passing.py
Pattern Category: 08 State Management
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 08.02: Context Passing Shared context between agents

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class ContextPassingState(TypedDict):
    """State for Context Passing pattern."""
    input: str
    output: str


def agent_function(state: ContextPassingState) -> ContextPassingState:

ContextPassingState: State for Context Passing pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Context Passing pattern: Stores prompt = f”Process this using Context Passing pattern – def create_context_passing(): Stores def create context passing()

Functions

`agent_function()`

Line: 17
Arguments: state

Context Passing agent implementation.

def agent_function(state: ContextPassingState) -> ContextPassingState:
    """Context Passing agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Context Passing pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_context_passing():

`create_context_passing()`

Line: 26
Arguments: “

Create Context Passing workflow.

def create_context_passing():
    """Create Context Passing workflow."""
    workflow = StateGraph(ContextPassingState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class ContextPassingState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Context Passing pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: ContextPassingState) -> ContextPassingState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Context Passing agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_context_passing definition

def create_context_passing():

Defines the create_context_passing function that processes the workflow state.

Line 27: Module docstring

    """Create Context Passing workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_context_passing()
    result = workflow.invoke({"input": "Test Context Passing", "output": ""})
    print(f"\n=== Context Passing Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: State Machine Pattern


    """
Pattern 08.01: State Machine
Explicit states and transitions
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class StateMachineState(TypedDict):
    """State for State Machine pattern."""
    input: str
    output: str


def agent_function(state: StateMachineState) -> StateMachineState:
    """State Machine agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using State Machine pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_state_machine():
    """Create State Machine workflow."""
    workflow = StateGraph(StateMachineState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_state_machine()
    result = workflow.invoke({"input": "Test State Machine", "output": ""})
    print(f"\n=== State Machine Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: State Machine Pattern

01 State Machine Pattern – Detailed Documentation

Overview

File: 08_state_management/01_state_machine.py
Pattern Category: 08 State Management
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 08.01: State Machine Explicit states and transitions

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class StateMachineState(TypedDict):
    """State for State Machine pattern."""
    input: str
    output: str


def agent_function(state: StateMachineState) -> StateMachineState:

StateMachineState: State for State Machine pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using State Machine pattern: Stores prompt = f”Process this using State Machine pattern – def create_state_machine(): Stores def create state machine()

Functions

`agent_function()`

Line: 17
Arguments: state

State Machine agent implementation.

def agent_function(state: StateMachineState) -> StateMachineState:
    """State Machine agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using State Machine pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_state_machine():

`create_state_machine()`

Line: 26
Arguments: “

Create State Machine workflow.

def create_state_machine():
    """Create State Machine workflow."""
    workflow = StateGraph(StateMachineState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class StateMachineState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for State Machine pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: StateMachineState) -> StateMachineState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """State Machine agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_state_machine definition

def create_state_machine():

Defines the create_state_machine function that processes the workflow state.

Line 27: Module docstring

    """Create State Machine workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_state_machine()
    result = workflow.invoke({"input": "Test State Machine", "output": ""})
    print(f"\n=== State Machine Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Adaptive Fallback


    """
Pattern 07.04: Adaptive Fallback
Dynamic fallback chain based on context
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class AdaptiveFallbackState(TypedDict):
    """State for Adaptive Fallback pattern."""
    input: str
    output: str


def agent_function(state: AdaptiveFallbackState) -> AdaptiveFallbackState:
    """Adaptive Fallback agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Adaptive Fallback pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_adaptive_fallback():
    """Create Adaptive Fallback workflow."""
    workflow = StateGraph(AdaptiveFallbackState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_adaptive_fallback()
    result = workflow.invoke({"input": "Test Adaptive Fallback", "output": ""})
    print(f"\n=== Adaptive Fallback Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Adaptive Fallback

04 Adaptive Fallback Pattern – Detailed Documentation

Overview

File: 07_adaptive/04_adaptive_fallback.py
Pattern Category: 07 Adaptive
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 07.04: Adaptive Fallback Dynamic fallback chain based on context

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class AdaptiveFallbackState(TypedDict):
    """State for Adaptive Fallback pattern."""
    input: str
    output: str


def agent_function(state: AdaptiveFallbackState) -> AdaptiveFallbackState:

AdaptiveFallbackState: State for Adaptive Fallback pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Adaptive Fallback pattern: Stores prompt = f”Process this using Adaptive Fallback pattern – def create_adaptive_fallback(): Stores def create adaptive fallback()

Functions

`agent_function()`

Line: 17
Arguments: state

Adaptive Fallback agent implementation.

def agent_function(state: AdaptiveFallbackState) -> AdaptiveFallbackState:
    """Adaptive Fallback agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Adaptive Fallback pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_adaptive_fallback():

`create_adaptive_fallback()`

Line: 26
Arguments: “

Create Adaptive Fallback workflow.

def create_adaptive_fallback():
    """Create Adaptive Fallback workflow."""
    workflow = StateGraph(AdaptiveFallbackState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class AdaptiveFallbackState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Adaptive Fallback pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: AdaptiveFallbackState) -> AdaptiveFallbackState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Adaptive Fallback agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_adaptive_fallback definition

def create_adaptive_fallback():

Defines the create_adaptive_fallback function that processes the workflow state.

Line 27: Module docstring

    """Create Adaptive Fallback workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_adaptive_fallback()
    result = workflow.invoke({"input": "Test Adaptive Fallback", "output": ""})
    print(f"\n=== Adaptive Fallback Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: A/B Testing Pattern


    """
Pattern 07.03: AB Testing
Parallel execution of alternatives with comparison
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class AbTestingState(TypedDict):
    """State for AB Testing pattern."""
    input: str
    output: str


def agent_function(state: AbTestingState) -> AbTestingState:
    """AB Testing agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using AB Testing pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_ab_testing():
    """Create AB Testing workflow."""
    workflow = StateGraph(AbTestingState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_ab_testing()
    result = workflow.invoke({"input": "Test AB Testing", "output": ""})
    print(f"\n=== AB Testing Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: A/B Testing Pattern

03 Ab Testing Pattern – Detailed Documentation

Overview

File: 07_adaptive/03_ab_testing.py
Pattern Category: 07 Adaptive
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 07.03: AB Testing Parallel execution of alternatives with comparison

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class AbTestingState(TypedDict):
    """State for AB Testing pattern."""
    input: str
    output: str


def agent_function(state: AbTestingState) -> AbTestingState:

AbTestingState: State for AB Testing pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using AB Testing pattern: Stores prompt = f”Process this using AB Testing pattern – def create_ab_testing(): Stores def create ab testing()

Functions

`agent_function()`

Line: 17
Arguments: state

AB Testing agent implementation.

def agent_function(state: AbTestingState) -> AbTestingState:
    """AB Testing agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using AB Testing pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_ab_testing():

`create_ab_testing()`

Line: 26
Arguments: “

Create AB Testing workflow.

def create_ab_testing():
    """Create AB Testing workflow."""
    workflow = StateGraph(AbTestingState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class AbTestingState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for AB Testing pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: AbTestingState) -> AbTestingState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """AB Testing agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_ab_testing definition

def create_ab_testing():

Defines the create_ab_testing function that processes the workflow state.

Line 27: Module docstring

    """Create AB Testing workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_ab_testing()
    result = workflow.invoke({"input": "Test AB Testing", "output": ""})
    print(f"\n=== AB Testing Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025

Agentic Orchestration Pattern: Learning Pattern


    """
Pattern 07.02: Learning
Agents improve from experience
"""
from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage


class LearningState(TypedDict):
    """State for Learning pattern."""
    input: str
    output: str


def agent_function(state: LearningState) -> LearningState:
    """Learning agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Learning pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_learning():
    """Create Learning workflow."""
    workflow = StateGraph(LearningState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":
    workflow = create_learning()
    result = workflow.invoke({"input": "Test Learning", "output": ""})
    print(f"\n=== Learning Pattern ===\nOutput: {result['output'][:200]}...")

Agentic Orchestration Pattern: Learning Pattern

02 Learning Pattern – Detailed Documentation

Overview

File: 07_adaptive/02_learning.py
Pattern Category: 07 Adaptive
Implementation: LangGraph/CrewAI

Pattern Description

Pattern 07.02: Learning Agents improve from experience

Code Structure

Imports

from typing import TypedDict
from langgraph.graph import StateGraph, END
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage

State Definition

class LearningState(TypedDict):
    """State for Learning pattern."""
    input: str
    output: str


def agent_function(state: LearningState) -> LearningState:

LearningState: State for Learning pattern.

State Fields: – input: Stores input – output: Stores output – def agent_function(state: Stores def agent function(state – prompt = f"Process this using Learning pattern: Stores prompt = f”Process this using Learning pattern – def create_learning(): Stores def create learning()

Functions

`agent_function()`

Line: 17
Arguments: state

Learning agent implementation.

def agent_function(state: LearningState) -> LearningState:
    """Learning agent implementation."""
    llm = ChatOpenAI()
    prompt = f"Process this using Learning pattern: {state['input']}"
    response = llm.invoke([HumanMessage(content=prompt)])
    state["output"] = str(response.content)
    return state


def create_learning():

`create_learning()`

Line: 26
Arguments: “

Create Learning workflow.

def create_learning():
    """Create Learning workflow."""
    workflow = StateGraph(LearningState)
    workflow.add_node("agent", agent_function)
    workflow.set_entry_point("agent")
    workflow.add_edge("agent", END)
    return workflow.compile()


if __name__ == "__main__":

Line-by-Line Explanation

Line 5: Import statement

from typing import TypedDict

Imports TypedDict for creating type-safe state dictionaries.

Line 6: Import statement

from langgraph.graph import StateGraph, END

Imports StateGraph – the core LangGraph class for building workflows.

Line 7: Import statement

from langchain_openai import ChatOpenAI

Imports ChatOpenAI for LLM interactions.

Line 8: Import statement

from langchain_core.messages import HumanMessage

Imports HumanMessage for formatting prompts to the LLM.

Line 11: Class definition

class LearningState(TypedDict):

Defines the state structure using TypedDict for type safety.

Line 12: Module docstring

    """State for Learning pattern."""

Describes the pattern and its purpose.

Line 17: Function agent_function definition

def agent_function(state: LearningState) -> LearningState:

Defines the agent_function function that processes the workflow state.

Line 18: Module docstring

    """Learning agent implementation."""

Describes the pattern and its purpose.

Line 26: Function create_learning definition

def create_learning():

Defines the create_learning function that processes the workflow state.

Line 27: Module docstring

    """Create Learning workflow."""

Describes the pattern and its purpose.

Usage Example

if __name__ == "__main__":
    workflow = create_learning()
    result = workflow.invoke({"input": "Test Learning", "output": ""})
    print(f"\n=== Learning Pattern ===\nOutput: {result['output'][:200]}...")

Key Concepts

State Management: Uses TypedDict to maintain workflow state
Graph-Based Flow: LangGraph orchestrates execution flow
Agent Functions: Each node processes and updates state
LLM Integration: Uses ChatOpenAI for intelligent processing

Pattern Benefits

✅ Type-safe state management
✅ Clear workflow visualization
✅ Modular and reusable components
✅ Easy to test and debug
✅ Scalable architecture

Related Patterns

See other patterns in the same category
Explore combination possibilities
Check the main documentation for pattern relationships

November 4, 2025